Skip to main content

Table 4 Relationships between the identified and manual wheat ear counting for the three data sets

From: Wheat ear counting using K-means clustering segmentation and convolutional neural network

Sites

Dataset

Samples size

Slope

Intercept

RMSE (ears)

R2

RRMSE (%)

Bias (ears)

Xuchang

Test

60

0.97

8.53

0.03

0.97

0.01

1.40

May 6

10

0.88

41.25

22.54

0.82

9.43

6.90

May 14

10

0.82

47.97

16.10

0.93

8.26

11.60

May 16

10

0.85

46.10

10.30

0.84

3.54

2.90

May 20

10

0.91

21.79

3.24

0.99

1.39

0.10

All dates

40

0.88

34.68

14.87

0.92

6.21

7.10

UAV

20

0.87

18.52

9.47

0.97

5.40

-5.00

All

120

0.97

9.05

10.84

0.96

4.63

2.24

  1. The K-means-CNN counting was highly correlated with manual counting